Guidance for Industry
Providing Clinical Evidence of
Effectiveness for Human Drug and
Biological Products
U.S. Department of Health and Human Services
Food and Drug Administration
Center for Drug Evaluation and Research (CDER)
Center for Biologics Evaluation and Research (CBER)
May 1998
Clinical 6
Guidance for Industry
Providing Clinical Evidence of
Effectiveness for Human Drugs and
Biological Products
Additional copies are available from:
the Drug Information Branch (HFD-210),
Center for Drug Evaluation and Research (CDER),
5600 Fishers Lane, Rockville, MD 20857 (Tel) 301-827-4573
Internet at http://www.fda.gov/cder/guidance/index.htm
or
Office of Communication,
Training, and Manufacturers Assistance (HFM-40)
Center for Biologics Evaluation and Research (CBER)
1401 Rockville Pike, Rockville, MD 20852-1448
http://www.fda.gov/cber/guidelines.htm
(Fax) 888-CBERFAX or 301-827-3844
(Voice Information) 800-835-4709 or 301-827-1800
U.S. Department of Health and Human Services
Food and Drug Administration
Center for Drug Evaluation and Research (CDER)
Center for Biologics Evaluation and Research (CBER)
May 1998
Clinical 6
TABLE OF CONTENTS
I.
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
II.
QUANTITY OF EVIDENCE NECESSARY TO SUPPORT EFFECTIVENESS . . . . . 2
III.
A.
Legal Standards for Drug and Biological Products . . . . . . . . . . . . . . . . . . . . . . . 2
B.
Scientific Basis for the Legal Standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
C.
The Quantity of Evidence to Support Effectiveness . . . . . . . . . . . . . . . . . . . . . . 6
DOCUMENTATION OF THE QUALITY OF EVIDENCE SUPPORTING AN
EFFECTIVENESS CLAIM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
A.
Reliance on Less Than Usual Access to Clinical Data
or Detailed Study Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
B.
Reliance on Studies with Alternative, Less Intensive
Quality Control/On-Site Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
GUIDANCE FOR INDUSTRY1
Providing Clinical Evidence of
Effectiveness2 for Human Drug and Biological Products
I.
INTRODUCTION
This document is intended to provide guidance to applicants planning to file new drug
applications (NDAs), biologics license applications (BLAs), or applications for supplemental
indications on the evidence to be provided to demonstrate effectiveness.
This document is also intended to meet the requirements of subsections 403(b)(1) and (2) of the
Food and Drug Administration Modernization Act (the Modernization Act) of 1997 for human
drug and biological products (P.L. 105-115).3 Subsection 403(b)(1) directs FDA to provide
guidance on the circumstances in which published matter may be the basis for approval of a
supplemental application for a new indication. Section III of this guidance satisfies this
requirement by describing circumstances in which published matter may partially or entirely
support approval of a supplemental application. Subsection 403(b)(2) directs FDA to provide
guidance on data requirements that will avoid duplication of previously submitted data by
recognizing the availability of data previously submitted in support of an original application to
support approval of a supplemental application. Section II of this guidance satisfies this
requirement by describing a range of circumstances in which related existing data, whether from
an original application or other sources, may be used to support approval of a supplemental
application.
In 1962, Congress amended the Federal Food, Drug, and Cosmetic Act to add a requirement that,
to obtain marketing approval, manufacturers demonstrate the effectiveness of their products
through the conduct of adequate and well-controlled studies. Since then, the issue of what
constitutes sufficient evidence of effectiveness has been debated by the Agency, the scientific
community, industry, and others. Sound evidence of effectiveness is a crucial component of the
Agency’s benefit-risk assessment of a new product or use. At the same time, the demonstration
of effectiveness represents a major component of drug development time and cost; the amount
1
This guidance document represents the agency’s current thinking on providing clinical evidence of
effectiveness for human drug and biological products. It does not create or confer any rights for or on any person and
does not operate to bind FDA or the public. An alternative approach may be used if such approach satisfies the
requirements of the applicable statute, regulations, or both.
2
As used in this guidance, the term efficacy refers to the findings in an adequate and well-controlled clinical
trial or the intent of conducting such a trial and the term effectiveness refers to the regulatory determination that is made
on the basis of clinical efficacy and other data.
3
The Modernization Act requirements in Section 403 also apply to animal drugs and medical devices. These
products will be addressed in separate guidances.
and nature of the evidence needed can therefore be an important determinant of when and
whether new therapies become available to the public. The public health is best served by the
development of sound evidence of effectiveness in an efficient manner.
The science and practice of drug development and clinical evaluation have evolved significantly
since the effectiveness requirement for drugs was established, and this evolution has implications
for the amount and type of data needed to support effectiveness in certain cases. As a result of
medical advances in the understanding of pathogenesis and disease staging, it is increasingly likely
that clinical studies of drugs will be more narrowly defined to focus, for example, on a more
specific disease stage or clinically distinct subpopulation. As a consequence, product indications
are often narrower, the universe of possible indications is larger, and data may be available from a
number of studies of a drug in closely related indications that bear on a determination of its
effectiveness for a new use. Similarly, there may be studies of a drug in different populations,
studies of a drug alone or in combination, and studies of different doses and dosage forms, all of
which may support a particular new use of a drug. At the same time, progress in clinical
evaluation and clinical pharmacology have resulted in more rigorously designed and conducted
clinical efficacy trials, which are ordinarily conducted at more than one clinical site. This added
rigor and scope has implications for a study’s reliability, generalizability, and capacity to
substantiate effectiveness.
Given this evolution, the Agency has determined that it would be appropriate to articulate its
current thinking concerning the quantitative and qualitative standards for demonstrating
effectiveness of drugs and biologics. FDA hopes that this guidance will enable sponsors to plan
drug development programs that are sufficient to establish effectiveness without being excessive
in scope. The guidance should also bring greater consistency and predictability to FDA’s
assessment of the clinical trial data needed to support drug effectiveness.
Another major goal of this guidance is to encourage the submission of supplemental applications
to add new uses to the labeling of approved drugs. By articulating how it currently views the
quantity and quality of evidence necessary to support approval of a new use of a drug, FDA hopes
to illustrate that the submission of supplements for new uses need not be unduly burdensome.
II.
QUANTITY OF EVIDENCE NECESSARY TO SUPPORT EFFECTIVENESS
A.
Legal Standards for Drug and Biological Products
Drugs: The effectiveness requirement for drug approval was added to the Federal Food,
Drug, and Cosmetic Act (the Act or the FDC Act) in 1962. Between passage of the Act
in 1938 and the 1962 amendments, drug manufacturers were required to show only that
their drugs were safe. The original impetus for the effectiveness requirement was
Congress's growing concern about the misleading and unsupported claims being made by
pharmaceutical companies about their drug products coupled with high drug prices. After
two years of hearings on these issues, Congress adopted the 1962 Drug Amendments,
2
which included a provision requiring manufacturers of drug products to establish a drug’s
effectiveness by "substantial evidence." Substantial evidence was defined in section
505(d) of the Act as “evidence consisting of adequate and well-controlled investigations,
including clinical investigations, by experts qualified by scientific training and experience
to evaluate the effectiveness of the drug involved, on the basis of which it could fairly and
responsibly be concluded by such experts that the drug will have the effect it purports or is
represented to have under the conditions of use prescribed, recommended, or suggested in
the labeling or proposed labeling thereof.”
Since the 1962 Amendments added this provision to the statute, discussions have ensued
regarding the quantity and quality of the evidence needed to establish effectiveness. With
regard to quantity, it has been FDA's position that Congress generally intended to require
at least two adequate and well-controlled studies, each convincing on its own, to establish
effectiveness. (See e.g., Final Decision on Benylin, 44 FR 51512, 518 (August 31, 1979);
Warner-Lambert Co. V. Heckler, 787 F. 2d 147 (3d Cir. 1986)). FDA’s position is based
on the language in the statute4 and the legislative history of the 1962 amendments.
Language in a Senate report suggested that the phrase "adequate and well-controlled
investigations" was designed not only to describe the quality of the required data but the
"quantum" of required evidence. (S. Rep. No. 1744, Part 2, 87th Cong. 2d Sess. 6
(1962))
Nevertheless, FDA has been flexible within the limits imposed by the congressional
scheme, broadly interpreting the statutory requirements to the extent possible where the
data on a particular drug were convincing. In some cases, FDA has relied on pertinent
information from other adequate and well-controlled studies of a drug, such as studies of
other doses and regimens, of other dosage forms, in other stages of disease, in other
populations, and of different endpoints, to support a single adequate and well-controlled
study demonstrating effectiveness of a new use. In these cases, although there is only one
study of the exact new use, there are, in fact, multiple studies supporting the new use, and
expert judgment could conclude that the studies together represent substantial evidence of
effectiveness. In other cases, FDA has relied on only a single adequate and wellcontrolled efficacy study to support approval — generally only in cases in which a single
multicenter study of excellent design provided highly reliable and statistically strong
evidence of an important clinical benefit, such as an effect on survival, and a confirmatory
study would have been difficult to conduct on ethical grounds.
In section 115(a) of the Modernization Act, Congress amended section 505(d) of the Act
to make it clear that the Agency may consider “data from one adequate and wellcontrolled clinical investigation and confirmatory evidence” to constitute substantial
4
Section 505(d) of the Act uses the plural form in defining “substantial evidence” as “adequate and wellcontrolled investigations, including clinical investigations.” See also use of “investigations” in section 505(b) of the
Act, which lists the contents of a new drug application.
3
evidence if FDA determines that such data and evidence are sufficient to establish
effectiveness. In making this clarification, Congress confirmed FDA’s interpretation of the
statutory requirements for approval and acknowledged the Agency’s position that there
has been substantial progress in the science of drug development resulting in higher quality
clinical trial data.
Biologics.
Biological products are approved under authority of section 351 of the
Public Health Service Act (PHS Act) (42 U.S.C.§ 262). Under section 351, as in effect
since 1944, licenses for biologics have been issued only upon a showing that the products
meet standards designed to ensure the “continued safety, purity, and potency” of the
products. Potency has long been interpreted to include effectiveness (21 CFR 600.3(s)).
In 1972, FDA initiated a review of the safety and effectiveness of all previously licensed
biologics. The Agency stated then that proof of effectiveness would consist of controlled
clinical investigations as defined in the provision for “adequate and well-controlled
studies” for new drugs (21 CFR 314.126), unless waived as not applicable to the
biological product or essential to the validity of the study when an alternative method is
adequate to substantiate effectiveness (21 CFR 601.25 (d) (2)). One such adequate
alternative was identified to be serological response data where a previously accepted
correlation with clinical effectiveness exists. As with nonbiological drug products, FDA
has approved biological products based on single, multicenter studies with strong results.
Although section 123(a) of the Modernization Act amended section 351 of the PHS Act
to make it clear that separate licenses are not required for biological products and the
establishments at which the products are made, the evidentiary standard for a biological
product was not changed: the product must be shown to be “safe, pure, and potent”
(section 351 (a)(2) of the PHS Act as amended). In the Modernization Act (section
123(f)) Congress also directed the agency to take measures to “minimize differences in the
review and approval” of products required to have approved BLAs under section 351 of
the PHS Act and products required to have approved NDAs under section 505(b)(1) of
the FDC Act.
B.
Scientific Basis for the Legal Standard
The usual requirement for more than one adequate and well-controlled investigation
reflects the need for independent substantiation of experimental results. A single clinical
experimental finding of efficacy, unsupported by other independent evidence, has not
usually been considered adequate scientific support for a conclusion of effectiveness. The
reasons for this include the following.
!
Any clinical trial may be subject to unanticipated, undetected, systematic biases.
These biases may operate despite the best intentions of sponsors and investigators,
and may lead to flawed conclusions. In addition, some investigators may bring
conscious biases to evaluations.
4
!
The inherent variability in biological systems may produce a positive trial result by
chance alone. This possibility is acknowledged, and quantified to some extent, in
the statistical evaluation of the result of a single efficacy trial. It should be noted,
however, that hundreds of randomized clinical efficacy trials are conducted each
year with the intent of submitting favorable results to FDA. Even if all drugs tested
in such trials were ineffective, one would expect one in forty of those trials to
“demonstrate” efficacy by chance alone at conventional levels of statistical
significance.5 It is probable, therefore, that false positive findings (i.e., the chance
appearance of efficacy with an ineffective drug) will occur and be submitted to
FDA as evidence of effectiveness. Independent substantiation of a favorable result
protects against the possibility that a chance occurrence in a single study will lead
to an erroneous conclusion that a treatment is effective.
!
Results obtained in a single center may be dependent on site or investigator
specific factors (e.g., disease definition, concomitant treatment, diet). In such
cases, the results, although correct, may not be generalizable to the intended
population. This possibility is the primary basis for emphasizing the need for
independence in substantiating studies.
!
Rarely, favorable efficacy results are the product of scientific fraud.
Although there are statistical, methodologic, and other safeguards to address the identified
problems, they are often inadequate to address these problems in a single trial.
Independent substantiation of experimental results addresses such problems by providing
consistency across more than one study, thus greatly reducing the possibility that a biased,
chance, site-specific, or fraudulent result will lead to an erroneous conclusion that a drug
is effective.
The need for independent substantiation has often been referred to as the need for
replication of the finding. Replication may not be the best term, however, as it may imply
that precise repetition of the same experiment in other patients by other investigators is the
only means to substantiate a conclusion. Precise replication of a trial is only one of a
number of possible means of obtaining independent substantiation of a clinical finding and,
at times, can be less than optimal as it could leave the conclusions vulnerable to any
systematic biases inherent to the particular study design. Results that are obtained from
studies that are of different design and independent in execution, perhaps evaluating
different populations, endpoints, or dosage forms, may provide support for a conclusion of
effectiveness that is as convincing as, or more convincing than, a repetition of the same
study.
5
p-value = 0.05, two-tailed, which implies an error rate in the efficacy (false positive) tail of 0.025 or one in
forty.
5
C.
The Quantity of Evidence to Support Effectiveness
The following three sections provide guidance on the quantity of evidence needed in
particular circumstances to establish substantial evidence of effectiveness. Section 1
addresses situations in which effectiveness of a new use may be extrapolated entirely from
existing efficacy studies. Section 2 addresses situations in which a single adequate and
well-controlled study of a specific new use can be supported by information from other
related adequate and well-controlled studies, such as studies in other phases of a disease,
in closely related diseases, of other conditions of use (different dose, duration of use,
regimen), of different dosage forms, or of different endpoints. Section 3 addresses
situations in which a single multicenter study, without supporting information from other
adequate and well-controlled studies, may provide evidence that a use is effective.
In each of these situations, it is assumed that any studies relied on to support effectiveness
meet the requirements for adequate and well-controlled studies in 21 CFR 314.126. It
should also be appreciated that reliance on a single study of a given use, whether alone or
with substantiation from related trial data, leaves little room for study imperfections or
contradictory (nonsupportive) information. In all cases, it is presumed that the single
study has been appropriately designed, that the possibility of bias due to baseline
imbalance, unblinding, post-hoc changes in analysis, or other factors is judged to be
minimal, and that the results reflect a clear prior hypothesis documented in the protocol.
Moreover, a single favorable study among several similar attempts that failed to support a
finding of effectiveness would not constitute persuasive support for a product use unless
there were a strong argument for discounting the outcomes in the studies that failed to
show effectiveness (e.g., study obviously inadequately powered or lack of assay sensitivity
as demonstrated in a three-arm study by failure of the study to show efficacy of a known
active agent).
Whether to rely on a single study to support an effectiveness determination is not often an
issue in contemporary drug development. In most drug development situations, the need
to find an appropriate dose, to study patients of greater and lesser complexity or severity
of disease, to compare the drug to other therapy, to study an adequate number of patients
for safety purposes, and to otherwise know what needs to be known about a drug before it
is marketed will result in more than one adequate and well-controlled study upon which to
base an effectiveness determination.
This guidance is not intended to provide a complete listing of the circumstances in which
existing efficacy data may provide independent substantiation of related claims; rather, it
provides examples of the reasoning that may be employed. The examples are applicable
whether the claim arises in the original filing of an NDA or BLA, or in a supplemental
application.
6
1.
Extrapolation from Existing Studies
In certain cases, effectiveness of an approved drug product for a new indication, or
effectiveness of a new product, may be adequately demonstrated without
additional adequate and well-controlled clinical efficacy trials. Ordinarily, this will
be because other types of data provide a way to apply the known effectiveness to a
new population or a different dose, regimen or dosage form. The following are
examples of situations in which effectiveness might be extrapolated from efficacy
data for another claim or product.
a.
Pediatric uses
The rule revising the Pediatric Use section of product labeling (21 CFR
201.57(f)(9)(iv)) makes allowance for inclusion of pediatric use
information in labeling without controlled clinical trials of the use in
children. In such cases, a sponsor must provide other information to
support pediatric use, and the Agency must conclude that the course of the
disease and the effects of the drug are sufficiently similar in the pediatric
and adult populations to permit extrapolation from adult efficacy data to
pediatric patients. Evidence that could support a conclusion of similar
disease course and similar drug effect in adult and pediatric populations
includes evidence of common pathophysiology and natural history of the
disease in the adult and pediatric populations, evidence of common drug
metabolism and similar concentration-response relationships in each
population, and experience with the drug, or other drugs in its therapeutic
class, in the disease or condition or related diseases or conditions.
Examples in which pediatric use labeling information has been extrapolated
from adult efficacy data include ibuprofen for pain and loratidine for
seasonal allergic rhinitis.
b.
Bioequivalence
The effectiveness of alternative formulations and new dosage strengths may
be assessed on the basis of evidence of bioequivalence.
c.
Modified-release dosage forms
In some cases, modified release dosage forms may be approved on the
basis of pharmacokinetic data linking the new dosage form to a previously
studied immediate-release dosage form. Because the pharmacokinetic
patterns of modified-release and immediate-release dosage forms are not
identical, it is generally important to have some understanding of the
relationship of blood concentration to response, including an understanding
of the time course of that relationship, to extrapolate the immediate-release
7
data to the modified-release dosage form.
d.
Different doses, regimens, or dosage forms
Dose-response relationships are generally continuous such that information
about the effectiveness of one dose, dosage regimen, or dosage form is
relevant to the effectiveness of other doses, regimens, or dosage forms.
Where blood levels and exposure are not very different, it may be possible
to conclude that a new dose, regimen, or dosage form is effective on the
basis of pharmacokinetic data alone. Even if blood levels are quite
different, if there is a well-understood relationship between blood
concentration and response, including an understanding of the time course
of that relationship, it may be possible to conclude that a new dose,
regimen, or dosage form is effective on the basis of pharmacokinetic data
without an additional clinical efficacy trial. In this situation,
pharmacokinetic data, together with the well-defined
pharmacokinetic/pharmacodynamic (PK/PD) relationship, are used to
translate the controlled trial results from one dose, regimen, or dosage
form to a new dose, regimen, or dosage form (See also section II.C.2.a).
2.
Demonstration of Effectiveness by a Single Study of a New Use, with
Independent Substantiation From Related Study Data
The discussion that follows describes specific examples in which a single study of a
new use, with independent substantiation from study data in related uses, could
provide evidence of effectiveness. In these cases, the study in the new use and the
related studies support the conclusion that the drug has the effect it is purported to
have. Whether related studies are capable of substantiating a single
study of a new use is a matter of judgment and depends on the quality and
outcomes of the studies and the degree of relatedness to the new use.
a.
Different doses, regimens, or dosage forms
As discussed in Sections II.C.1.d, it may be possible to conclude that a new
dose, regimen, or dosage form is effective on the basis of pharmacokinetic
data without an additional clinical efficacy trial where blood levels and
exposure are not very different or, even if quite different, there is a wellunderstood relationship between blood concentration and response. Where
the relationship between blood concentration and response is not so well
understood and the pharmacokinetics of the new dose, regimen, or dosage
form differ from the previous one, clinical efficacy data will likely be
necessary to support effectiveness of a new regimen. In this case, a single
additional efficacy study should ordinarily be sufficient. For example, a
single controlled trial was needed to support the recent approval of a once
8
daily dose of risperidone because the once daily and twice daily regimens
had different pharmacokinetics and risperidone’s PK/PD relationship was
not well understood.
b.
Studies in other phases of the disease
In many cases, therapies that are effective in one phase of a disease are
effective in other disease phases, although the magnitude of the benefit and
benefit-to-risk relationship may differ in these other phases. For example,
if a drug is known to be effective in patients with a refractory stage of a
particular cancer, a single adequate and well-controlled study of the drug in
an earlier stage of the same tumor will generally be sufficient evidence of
effectiveness to support the new use.
c.
Studies in other populations
Often, responses in subsets of a particular patient population are
qualitatively similar to those in the whole population. In most cases,
separate studies of effectiveness in demographic subsets are not needed
(see also discussion of the pediatric population in section II.C.1.a)
However, where further studies are needed, a single study would ordinarily
suffice to support effectiveness in age, race, gender, concomitant disease,
or other subsets for a drug already shown to be generally effective in a
condition or to be effective in one population. For example, a single study
was sufficient to support tamoxifen use in breast cancer in males.
d.
Studies in combination or as monotherapy
For a drug known to be effective as monotherapy, a single adequate and
well-controlled study is usually sufficient to support effectiveness of the
drug when combined with other therapy (as part of a multidrug regimen or
in a fixed-dose combination). Similarly, known effectiveness of a drug as
part of a combination (i.e., its contribution to the effect of the combination
is known) would usually permit reliance on a single study of appropriate
design to support its use as monotherapy, or as part of a different
combination, for the same use. For example, a single study of a new
combination vaccine designed to demonstrate adequate immune response
will ordinarily provide sufficient evidence of effectiveness if the new
combination contains products or antigens already proven to be effective
alone or in other combinations. These situations are common for
oncologic and antihypertensive drugs, but occur elsewhere as well.
9
e.
Studies in a closely related disease
Studies in etiologically or pathophysiologically related conditions, or
studies of a symptom common to several diseases (e.g., pain) can support
each other, allowing initial approval of several uses or allowing additional
claims based on a single adequate and well-controlled study. For example,
certain anti-coagulant or anti-platelet therapies could be approved for use
in two different settings based on individual studies in unstable
angina/acute coronary syndrome and in the postangioplasty state. Because
the endpoints studied and the theoretical basis for use of an anti-coagulant
or anti-platelet drug are similar, each study supports the other for each
claim. Similarly, single analgesic studies in several painful conditions
would ordinarily be sufficient to support either a general analgesic
indication or multiple specific indications. The recent approval of
lamotrigine for treatment of Lennox-Gastaut Syndrome (a rare, largely
pediatric, generalized seizure disorder) was based on a
single adequate and well-controlled trial, due in part to related data
showing efficacy of the drug in partial-onset seizures in adults.
f.
Studies in less closely related diseases, but where the general
purpose of therapy is similar
Certain classes of drug therapy, such as antimicrobials and antineoplastics,
are appropriate interventions across a range of different diseases. For
therapies of this type, evidence of effectiveness in one disease could
provide independent substantiation of effectiveness in a quite different
disease. For example, it is possible to argue that evidence of effectiveness
of an antimicrobial in one infectious disease setting may support reliance on
a single study showing effectiveness in other settings where the causative
pathogens, characteristics of the site of infection that affect the disease
process (e.g., structure and immunology) and patient population are
similar.6 Similarly, for an oncologic drug, evidence of effectiveness in one
or more tumor types may support reliance on a single study showing
effectiveness against a different kind of tumor, especially if the tumor types
have a common biological origin.
g.
Studies of different clinical endpoints
Demonstration of a beneficial effect in different studies on two different
clinically meaningful endpoints could cross-substantiate a claim for
6
See Division of Anti-Infective Drug Products: Points to Consider in the Clinical Development and Labeling
of Anti-Infective Drug Products, October 1992.
10
effectiveness for each outcome. For example, the initial claim for
effectiveness of enalapril for heart failure was supported by one study
showing symptom improvement over several months and a second study
showing improved survival in a more severely ill population. The two
different findings, each from an adequate and well-controlled study, led to
the conclusion that enalapril was effective in both treating symptoms and
improving survival.
h.
Pharmacologic/pathophysiologic endpoints
When the pathophysiology of a disease and the mechanism of action of a
therapy are very well understood, it may be possible to link specific
pharmacologic effects to a strong likelihood of clinical effectiveness. A
pharmacologic effect that is accepted as a validated surrogate endpoint can
support ordinary approval (e.g., blood pressure effects, cholesterollowering effects) and a pharmacologic effect that is considered reasonably
likely to predict clinical benefit can support accelerated approval under the
conditions described in 21 CFR 314 Subpart H and 21 CFR 601 Subpart E
(e.g., CD4 count and viral load effects to support effectiveness of anti-viral
drugs for HIV infection). When the pharmacologic effect is not considered
an acceptable effectiveness endpoint, but the linkage between it and the
clinical outcome is strong, not merely on theoretical grounds but based on
prior therapeutic experience or well-understood pathophysiology, a single
adequate and well-controlled study showing clinical efficacy can sometimes
be substantiated by persuasive data from a well-controlled study or studies
showing the related pharmacologic effect.
For example, a single clearly positive trial can be sufficient to support
approval of a replacement therapy such as a coagulation factor, when it is
combined with clear evidence that the condition being treated is caused by
a deficiency of that factor. Demonstration of physical replacement of the
deficient factor or restoration of the missing physiologic activity provides
strong substantiation of the clinical effect. The corrective treatment of an
inborn error of metabolism could be viewed similarly. In the case of
preventive vaccines, one adequate and well-controlled clinical trial may be
supported by compelling animal challenge/protection models, human
serological data, passive antibody data, or pathogenesis information. The
more evidence there is linking effects on the pharmacologic endpoint to
improvement or prevention of the disease, the more persuasive the
argument for reliance on a single clinical efficacy study.
Note, however, that plausible beneficial pharmacologic effects have often
not correlated with clinical benefit, and, therefore, caution must be
observed in relying on a pharmacologic effect as contributing to evidence
11
of effectiveness. For example, pharmacologic effects such as arrhythmia
suppression by Type 1 antiarrhythmics and increased cardiac output by
phosphodiesterase inhibitors or beta adrenergic inotropes resulted in
increased mortality, rather than, as was expected, decreased sudden death
and improved outcome in heart failure. The reasons for the absence of an
expected correlation between pharmacologic and clinical effects are diverse
and can include an incompletely understood relationship between the
pharmacologic effect and the clinical benefit and the presence of other
pharmacologic effects attributable to a drug in addition to the effect being
measured and thought to be beneficial. Generally, the utility of
pharmacologic outcomes in providing independent substantiation will be
greatest where there is prior experience with the pharmacologic class.
Even in this case, however, it is difficult to be certain that a pharmacologic
effect that correlates with a clinical benefit accounts for all the clinical
benefit or that other effects are not present and relevant.
3.
Evidence of Effectiveness from a Single Study
When the effectiveness requirement was originally implemented in 1962, the
prevailing efficacy study model was a single institution, single investigator,
relatively small trial with relatively loose blinding procedures, and little attention to
prospective study design and identification of outcomes and analyses. At present,
major clinical efficacy studies are typically multicentered, with clear, prospectively
determined clinical and statistical analytic criteria. These studies are less
vulnerable to certain biases, are often more generalizable, may achieve very
convincing statistical results, and can often be evaluated for internal consistency
across subgroups, centers, and multiple endpoints.
The added rigor and size of contemporary clinical trials have made it possible to
rely, in certain circumstances, on a single adequate and well-controlled study,
without independent substantiation from another controlled trial, as a sufficient
scientific and legal basis for approval. For example, the approval of timolol for
reduction of post-infarction mortality was based on a single, particularly persuasive
(low p-value), internally consistent, multicenter study that demonstrated a major
effect on mortality and reinfarction rate. For ethical reasons, the study was
considered unrepeatable. The Center for Biologics Evaluation and Research has
also approved a number of products based upon a single persuasive study. The
Agency provided a general statement in 1995 describing when a single, multicenter
study may suffice (60 FR 39181; August 1, 1995), but the Agency has not
comprehensively described the situations in which a single adequate and wellcontrolled study might be considered adequate support for an effectiveness claim,
or the characteristics of a single study that could make it adequate support for an
effectiveness claim.
12
Whether to rely on a single adequate and well-controlled study is inevitably a
matter of judgment. A conclusion based on two persuasive studies will always be
more secure than a conclusion based on a single, comparably persuasive study.
For this reason, reliance on only a single study will generally be limited to
situations in which a trial has demonstrated a clinically meaningful effect on
mortality, irreversible morbidity, or prevention of a disease with potentially serious
outcome and confirmation of the result in a second trial would be practically or
ethically impossible. For example, sequential repetition of strongly positive trials
that demonstrated a decrease in post-infarction mortality, prevention of
osteoporotic fractures, or prevention of pertussis would present significant ethical
concerns. Repetition of positive trials showing only symptomatic benefit would
generally not present the same ethical concerns.
The discussion that follows identifies the characteristics of a single adequate and
well-controlled study that could make the study adequate support for an
effectiveness claim. Although no one of these characteristics is necessarily
determinative, the presence of one or more in a study can contribute to a
conclusion that the study would be adequate to support an effectiveness claim.
a.
Large multicenter study
In a large multicenter study in which (1) no single study site provided an
unusually large fraction of the patients and (2) no single investigator or site
was disproportionately responsible for the favorable effect seen, the study’s
internal consistency lessens concerns about lack of generalizability of the
finding or an inexplicable result attributable only to the practice of a single
investigator. If analysis shows that a single
site is largely responsible for the effect, the credibility of a multicenter
study is diminished.
b.
Consistency across study subsets
Frequently, large trials have relatively broad entry criteria and the study
populations may be diverse with regard to important covariates such as
concomitant or prior therapy, disease stage, age, gender or race. Analysis
of the results of such trials for consistency across key patient subsets
addresses concerns about generalizability of findings to various populations
in a manner that may not be possible with smaller trials or trials with more
narrow entry criteria. For example, the timolol postinfarction study
randomized patients separately within three severity strata. The study
showed positive effects on survival in each stratum supporting a conclusion
that the drug’s utility was not limited to a particular disease stage (e.g.,
relatively low or high severity).
13
c.
Multiple studies in a single study
Properly designed factorial studies may be analyzed as a series of pairwise
comparisons, representing, within a single study, separate demonstrations
of activity of a drug as monotherapy and in combination with another drug.
This model was successfully used in ISIS II, which showed that for patients
with a myocardial infarction both aspirin and streptokinase had favorable
effects on survival when used alone and when combined (aspirin alone and
streptokinase alone were each superior to placebo; aspirin and
streptokinase in combination were superior to aspirin alone and to
streptokinase alone). This represented two separate (but not completely
independent) demonstrations of the effectiveness of aspirin and
streptokinase.
d.
Multiple endpoints involving different events
In some cases, a single study will include several important, prospectively
identified primary or secondary endpoints, each of which represents a
beneficial, but different, effect. Where a study shows statistically
persuasive evidence of an effect on more than one of such endpoints, the
internal weight of evidence of the study is enhanced. For example, the
approval of beta-interferon (Betaseron) for prevention of exacerbations in
multiple sclerosis was based on a single multicenter study, at least partly
because there were both a decreased rate of exacerbations and a decrease
in MRI-demonstrated disease activity — two entirely different, but
logically related, endpoints.
Similarly, favorable effects on both death and nonfatal myocardial
infarctions in a lipid-lowering, postangioplasty, or postinfarction study
would, in effect, represent different, but consistent, demonstrations of
effectiveness, greatly reducing the possibility that a finding of reduced
mortality was a chance occurrence. For example, approval of abciximab as
adjunctive treatment for patients undergoing complicated angioplasty or
atherectomy was supported by a single study with a strong overall result on
the combined endpoint (decreased the combined total of deaths, new
infarctions, and need for urgent interventions) and statistically significant
effects in separate evaluations of two components of the combined
endpoint (decreased new infarctions and decreased need for urgent
interventions). In contrast, a beneficial effect on multiple endpoints that
evaluate essentially the same phenomenon and correlate strongly, such as
mood change on two different depression scales or SGOT and CPK levels
postinfarction, does not significantly enhance the internal weight of the
evidence from a single trial.
14
Although two consistent findings within a single study usually provide
reassurance that a positive treatment effect is not due to chance, they do
not protect against bias in study conduct or biased analyses. For example,
a treatment assignment not well balanced for important prognostic
variables could lead to an apparent effect on both endpoints. Thus, close
scrutiny of study design and conduct are critical to evaluating this type of
study.
e.
Statistically very persuasive finding
In a multicenter study, a very low p-value indicates that the result is highly
inconsistent with the null hypothesis of no treatment effect. In some
studies it is possible to detect nominally statistically significant results in
data from several centers, but, even where that is not possible, an overall
extreme result and significance level means that most study centers had
similar findings. For example, the thrombolysis trials of streptokinase (ISIS
II, GISSI) had very sizable treatment effects and very low p-values, greatly
adding to their persuasiveness. Preventive vaccines for infectious
disease indications with a high efficacy rate (e.g., point estimate of efficacy
of 80% or higher and a reasonably narrow 95% confidence interval) have
been approved based on a single adequate and well-controlled trial.
4.
Reliance on a Single, Multicenter Study — Caveats
While acknowledging the persuasiveness of a single, internally consistent, strong
multicenter study, it must be appreciated that even a strong result can represent an
isolated or biased result, especially if that study is the only study suggesting
efficacy among similar studies. Recently, the apparent highly favorable effect of
vesnarinone, an inotropic agent, in heart failure (60% reduction of mortality in
what appeared to be a well-designed, placebo-controlled, multicenter trial with an
extreme p-value) has proven to be unrepeatable. In an attempt to substantiate the
finding, the same dose of the drug that seemed lifesaving in the earlier study
significantly increased mortality (by 26%), and a lower dose also appeared to have
a detrimental effect on survival. Although the population in the second study was,
on the whole, a sicker population than in the first, the outcomes in similarly sick
patients in each study were inconsistent so this factor does not explain the
contradictory results.
When considering whether to rely on a single multicenter trial, it is critical that the
possibility of an incorrect outcome be considered and that all the available data be
examined for their potential to either support or undercut reliance on a single
multicenter trial. In the case of vesnarinone, there were other data that were not
consistent with the dramatically favorable outcome in the multicenter study. These
data seemed to show an inverse dose-response relationship, showed no suggestion
15
of symptomatic benefit, and showed no effect on hemodynamic endpoints. These
inconsistencies led the Agency, with the advice
of its Cardio-Renal Advisory Committee, to refuse approval — a decision borne
out by the results of the subsequent study.
This example illustrates how inadequacies and inconsistencies in the data, such as
lack of pharmacologic rationale and lack of expected other effects accompanying a
critical outcome, can weaken the persuasiveness of a single trial. Although an
unexplained failure to substantiate the results of a favorable study in a second
controlled trial is not proof that the favorable study was in error — studies of
effective agents can fail to show efficacy for a variety of reasons — it is often
reason not to rely on the single favorable study.
III.
DOCUMENTATION OF THE QUALITY OF EVIDENCE SUPPORTING AN
EFFECTIVENESS CLAIM
When submitting the requisite quantity of data to support approval of a new product or new use
of an approved product, sponsors must also document that the studies were adequately designed
and conducted. Essential characteristics of adequate and well-controlled trials are described in
21 CFR 314.126. To demonstrate that a trial supporting an effectiveness claim is adequate and
well-controlled, extensive documentation of trial planning, protocols, conduct, and data handling
is usually submitted to the Agency, and detailed patient records are made available at the clinical
sites.
From a scientific standpoint, however, it is recognized that the extent of documentation necessary
depends on the particular study, the types of data involved, and the other evidence available to
support the claim. Therefore, the Agency is able to accept different levels of documentation of
data quality, as long as the adequacy of the scientific evidence can be assured. This section
discusses the factors that influence the extent of documentation needed, with particular emphasis
on studies evaluating new uses of approved drugs.
For the purposes of this section, the phrase documentation of the quality of evidence refers to (1)
the completeness of the documentation and (2) the ability to access the primary study data and the
original study-related records (e.g., subjects’ medical records, drug accountability records) for the
purposes of verifying the data submitted as evidence. These interrelated elements bear on a
determination of whether a study is adequate and well-controlled.
In practice, to achieve a high level of documentation, studies supporting claims are ordinarily
conducted in accordance with good clinical practices (GCPs). Sponsors routinely monitor all
clinical sites, and FDA routinely has access to the original clinical protocols, primary data, clinical
site source documents for on-site audits, and complete study reports.
16
However, situations often arise in which studies that evaluate the efficacy of a drug product lack
the full documentation described above (for example, full patient records may not be available) or
in which the study was conducted with less monitoring than is ordinarily seen in commercially
sponsored trials. Such situations are more common for supplemental indications because
postapproval studies are more likely to be conducted by parties other than the drug sponsor and
those parties may employ less extensive monitoring and data-gathering procedures than a sponsor.
Under certain circumstances, it is possible for sponsors to rely on such studies to support
effectiveness claims, despite less than usual documentation or monitoring. Some of those
circumstances are described below.
A.
Reliance on Less Than Usual Access to Clinical Data or Detailed Study
Reports
FDA’s access to primary data has proven to be important in many regulatory decisions.
There are also reasons to be skeptical of the conclusions of published reports of studies.
Experience has shown that such study reports do not always contain a complete, or
entirely accurate, representation of study plans, conduct and outcomes. Outright fraud
(i.e., deliberate deception) is unusual. However, incompleteness, lack of clarity,
unmentioned deviation from prospectively planned analyses, or an inadequate description
of how critical endpoint judgments or assessments were made are common flaws.
Typically, journal article peer reviewers only have access to a limited data set and
analyses, do not see the original protocol and amendments, may not know what happened
to study subjects that investigators determined to be non-evaluable, and thus may lack
sufficient information to detect critical omissions and problems. The utility of peer review
can also be affected by variability in the relevant experience and expertise of peer
reviewers. FDA's experiences with the Anturane Reinfarction Trial, as well as literature
reports of the efficacy of tacrine and the anti-sepsis HA-1A antibody, illustrate its
concerns with reliance on the published medical literature.
Notwithstanding these concerns, the presence of some of the factors discussed below can
make it possible for FDA to rely on studies for which it has less than usual access to data
or detailed study reports to partially or entirely (the so-called paper filing) support an
effectiveness claim. FDA’s reliance on a literature report to support an effectiveness claim
is more likely if FDA can obtain additional critical study details. Section 1 below
describes additional information that, if available, would increase the likelihood that a
study could be relied on to support an effectiveness claim. Section 2 describes factors that
may make efficacy findings sufficiently persuasive to permit reliance on the published
literature alone. Note that the factors outlined in Section 2 are relevant to an assessment
of the reliability of literature reports generally, whether alone, or accompanied by other
important information as discussed in Section 1.
1.
Submission of Published Literature or Other Reports in Conjunction with
Other Important Information that Enhances the Reliability of the Data
17
If a sponsor wishes to rely on a study conducted by another party and cannot
obtain the primary data from the study, for most well-conducted studies it is
possible to obtain other important information, such as a protocol documenting the
prospective plans for the trial, records of trial conduct and procedures, patient data
listings for important variables, and documentation of the statistical analysis. FDA
has considerable experience evaluating large multicenter outcome studies
sponsored by U.S. and European government agencies (NIH, British Medical
Research Council) and private organizations (the ISIS studies, the SAVE study)
for which there was limited access to primary study data, but for which other
critical information was available. Providing as many as possible of the following
important pieces of information about a study, in conjunction with the published
report, can increase the likelihood that the study can be relied on to support an
effectiveness claim:
a.
The protocol used for the study, as well as any important protocol
amendments that were implemented during the study and their relation to
study accrual or randomization.
b.
The prospective statistical analysis plan and any changes from the
original plan that occurred during or after the study, with particular note of
which analyses were performed pre- and post-unblinding.
c.
Randomization codes and documented study entry dates for the
subjects.
d.
Full accounting of all study subjects, including identification of any
subjects with on-treatment data who have been omitted from analysis and
the reasons for omissions, and an analysis of results using all subjects with
on-study data.
e.
Electronic or paper record of each subject’s data for critical
variables and pertinent baseline characteristics. Where individual subject
responses are a critical variable (e.g., objective responses in cancer
patients, clinical cures and microbial eradications in infectious disease
patients, death from a particular cause), detailed bases for the assessment,
such as the case report, hospital records, and narratives, should be
provided when possible.
f.
Where safety is a major issue, complete information for all deaths
and drop-outs due to toxicity. For postapproval supplemental uses,
however, there is generally less need for the results of lab tests or for
details of adverse event reports and, consequently, much more limited
documentation may be sufficient (e.g., only for unexpected deaths and
previously undescribed serious adverse effects). Exceptions to this
18
approach would include situations in which the population for the
supplemental use is so different that existing safety information has limited
application (e.g., thrombolysis in stroke patients versus myocardial
infarction patients) or where the new population presents serious safety
concerns (e.g., extension of a preventive vaccine indication from young
children to infants).
2.
Submission of Published Literature Reports Alone
The following factors increase the possibility of reliance on published reports alone
to support approval of a new product or new use:
a.
Multiple studies conducted by different investigators where each of
the studies clearly has an adequate design and where the findings across
studies are consistent.
b.
A high level of detail in the published reports, including clear and
adequate descriptions of statistical plans, analytic methods (prospectively
determined), and study endpoints, and a full accounting of all enrolled
patients.
c.
Clearly appropriate endpoints that can be objectively assessed and
are not dependent on investigator judgment (e.g., overall mortality, blood
pressure, or microbial eradication). Such endpoints are more readily
interpreted than more subjective endpoints such as cause-specific mortality
or relief of symptoms.
d.
Robust results achieved by protocol-specified analyses that yield a
consistent conclusion of efficacy and do not require selected post hoc
analyses such as covariate adjustment, subsetting, or reduced data sets
(e.g., analysis of only responders or compliant patients, or of an "eligible"
or “evaluable” subset).
e.
Conduct of studies by groups with properly documented operating
procedures and a history of implementing such procedures effectively.
There have been approvals based primarily or exclusively on published reports.
Examples include the initial approval of secretin for evaluation of pancreatic
function and recent approvals of bleomycin and talc for malignant pleural effusion
and doxycycline for malaria.
19
B.
Reliance on Studies with Alternative, Less Intensive Quality Control/On-Site
Monitoring
Industry-sponsored studies typically use extensive on-site and central monitoring and
auditing procedures to assure data quality. Studies supported by other sponsors may
employ less stringent procedures and may use no on-site monitoring at all. An
International Conference on Harmonisation guideline on good clinical practices,7 recently
accepted internationally, emphasizes that the extent of monitoring in a trial should be
based on trial-specific factors (e.g., design, complexity, size, and type of study outcome
measures) and that different degrees of on-site monitoring can be appropriate. In recent
years, many credible and valuable studies conducted by government or independent study
groups, often with important mortality outcomes, had very little on-site monitoring.
These studies have addressed quality control in other ways, such as by close control and
review of documentation and extensive guidance and planning efforts with investigators.
There is a long history of reliance on such studies for initial approval of drugs as well as
for additional indications. Factors that influence whether studies with limited or no
monitoring may be relied on include the following:
1.
The existence of a prospective plan to assure data quality.
2.
Studies that have features that make them inherently less susceptible to
bias, such as those with relatively simple procedures, noncritical entry criteria, and
readily assessed outcomes.
3.
The ability to sample critical data and make comparisons to supporting
records (e.g., hospital records).
4.
Conduct of the study by a group with established operating procedures and
a history of implementing such procedures effectively.
7
International Conference on Harmonisation Guidance for Industry E6, Good Clinical Practice: Consolidated
Guideline, April 1996.
20